VOGONS


First post, by dinth

User metadata
Rank Newbie
Rank
Newbie

Hi. I am trying to build a 486 computer, using my older motherboards and they all seem to be dead, but when i tried a third one is completely dead it started to feel unlikely.
For powering it i have bought a 150W ATX power supply (https://www.amazon.co.uk/gp/product/B09B2739V … e?ie=UTF8&psc=1), and a converter to AT. I hooked it up to a power supply tester and everything is all right.
I have dug out a couple of old graphic cards, one Virge and one ET3000.
I also purchased a tested Intel DX4/100 CPU online.
I have also bought an ISA POST tester (https://www.amazon.co.uk/gp/product/B07B65PBR … e?ie=UTF8&psc=1)
The first two motherboards i have tested had leaky batteries and some damage caused by them. All the caps looked all right though. After hooking them both looked completely dead - no beeps, no VGA output (i tried both cards with every mobo) and above anything else no POST codes on the ISA post tester.
The third motherboard has a ODIN encased battery, which is probably dead too, but no leaks outside and no acid damage anywhere. Caps looking all right too. I hooked everything up and unfortunately im getting same behaviour - no beeps, no VGA output and no POST codes ("----") on the ISA post tester. CPU was getting warm though. I had a look into the manual for the tester and it says that no codes are present when the CPU is dead. So i have purchased a new CPU on ebay, this time a tested AMD 100Mhz, replaced the CPU and still getting the same behaviour - the computer doesnt post but CPU is getting warm.
I dont know what is the model of the motherboard, but seems that CPU jumpers are currently set to 1x 50MHz. Not exactly sure about memory jumpers, but even if those are misconfigured, the motherboard should POST, right?
After powering the mobos, they are taking between 6 and 11W from the power supply, depending on the number of memory modules and if the GPU is inserted

I know that those mobos are super old, but the fact im getting the same issue (no POST, but with CPU powered) on three separate motherboards (which used to work) seems to be somewhat unlikely. Maybe im doing something wrong, or forgetting about something, it has been over 20 years since i last played with 486 gen computer? Could someone please sanity check me?

Last edited by dinth on 2022-11-18, 17:30. Edited 1 time in total.

Reply 2 of 18, by vstrakh

User metadata
Rank Member
Rank
Member

Did you close the switch that starts ATX psu (PS_ON# to GND)?
I don't remember how the Pico-PSU behaves, but I'd expect it to do the same as normal PSU does.
And I see some hanging wires with the clips that should go into the switch.

Reply 3 of 18, by gerwin

User metadata
Rank l33t
Rank
l33t
dinth wrote on 2022-11-18, 17:25:

I dont know what is the model of the motherboard

It is obviously very similar to this TK 8498F motherboard.

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 4 of 18, by dinth

User metadata
Rank Newbie
Rank
Newbie
vstrakh wrote on 2022-11-18, 17:39:

Did you close the switch that starts ATX psu (PS_ON# to GND)?
I don't remember how the Pico-PSU behaves, but I'd expect it to do the same as normal PSU does.
And I see some hanging wires with the clips that should go into the switch.

Yep I'm powering the mobo up by connecting those two wires

Reply 5 of 18, by clb

User metadata
Rank Newbie
Rank
Newbie

My 486 motherboard has a poorly programmed BIOS code that does not allow booting to DOS at all if the battery is out. It looked very much like that Odin battery. I replaced it with a new adapted one that allows placing a small button cell battery in the slot. ( Replacing the Dallas 12B887 )

Although when the battery was out, it was still possible to enter BIOS to modify settings, so the system did POST at least. Hmm, maybe that might not be the issue here.

Well, when you eventually do get it working, there may be some battery work ahead of you.

Reply 6 of 18, by paradigital

User metadata
Rank Member
Rank
Member

I’d double check all the CPU configuration jumpers and then check the voltage coming out of the regulator near the CPU socket.

I’ve had two 486 boards now with dead or dying voltage regulators.

Reply 7 of 18, by CoffeeOne

User metadata
Rank Oldbie
Rank
Oldbie
paradigital wrote on 2022-11-18, 18:08:

I’d double check all the CPU configuration jumpers and then check the voltage coming out of the regulator near the CPU socket.

I’ve had two 486 boards now with dead or dying voltage regulators.

I agree. Even if that would - most likely - not solve your problem: When testing 486 boards, it is good to have one 5V CPU. Like a normal Intel DX33 or an Intel DX2-66 with 5 volts.

Reply 8 of 18, by mkarcher

User metadata
Rank Oldbie
Rank
Oldbie
dinth wrote on 2022-11-18, 17:25:

I dont know what is the model of the motherboard, but seems that CPU jumpers are currently set to 1x 50MHz. Not exactly sure about memory jumpers, but even if those are misconfigured, the motherboard should POST, right?

Memory is usually not jumpered, cache is. Cache misjumpering doesn't prevent power-up. All DX4 CPUs (except special "overdrive" ones) are 3.3V CPUs. You should not power them with 5V. Rumor has it that 3.3V processors survive being run at 5V for some time, so probably running them at 5V doesn't kill them instantly, but you should make sure to only run them on 3.3V capable boards in 3.3V setting (or using a voltage adapter socket). It seems you got that one right, if the jumper manual for a similar board I found applies to your board.

You can not run a DX4 at x1, only at x2 or x3. The DX4 you have uses the "early AMD DX4" pinout, which is different from the Intel pinout. Use the jumpering for the "AMD DX/DX2/DX4" configuration, do not use the "Intel DX4" jumpering. The pin used to select x2 or x3 on the early AMD DX4 is different from the one used on later DX4 processors, but the manuals for similar boards indicate that the jumper scheme for AMD should run the DX4 at x3. So you are not running at 1x 50MHz, but you are running at 3x something. If it were 3x 50MHz, most AMD DX4/100 won't start, so the behaviour were expected. The photo you show in the first post has JP4/5/6 all closed, which selects 33MHz. That would result in 3*33MHz, which is correct for the AMD DX4. CPU jumpering on the photo seems to be correct for AMD processors, so the jumpering actually is correct, but it is not 1x 50MHz.

You said you have a POST card. Please verify that all voltage LEDs are on, the CLK LED is on, the RESET LED lights up for a short moment after power-up and then stays off. Remove RAM and the graphics card. The processor, the POST card and the board should be enough to get some POST codes. If the POST card stays on "----", make sure the processor is fully inserted to the socket.

Reply 9 of 18, by dinth

User metadata
Rank Newbie
Rank
Newbie

Thanks for the reply. I think i have identified the mobo to be very similar to this one - https://www.elhvb.com/webhq/models/486vlb3/mb1433ui.htm slightly different chipset and lack of KB0, but the jumpers are all in the same places.
Following this guide i set everything for Intel DX4/100 - https://www.elhvb.com/webhq/models/486vlb3/mb1433ui.txt.html (of course assuming that "1" on all of them is on the right hand side, as only few jumpers on the board show where the first pin is. Unfortunately nothing has changed, the post card shows ----.
It does have some extra functionality available after pressing the button on the card, but nothing regarding the button was mentioned in the card's manual (which was super brief and only listed post codes)

Regarding the LEDs on the card, im getting +5V, +12V and -12V. +3.5V is not lit. CLK is lit. Reset lits up shortly when powering up and also when connecting reset jumper.

Reply 10 of 18, by butjer1010

User metadata
Rank Newbie
Rank
Newbie
dinth wrote on 2022-11-18, 19:22:
Thanks for the reply. I think i have identified the mobo to be very similar to this one - https://www.elhvb.com/webhq/models/48 […]
Show full quote

Thanks for the reply. I think i have identified the mobo to be very similar to this one - https://www.elhvb.com/webhq/models/486vlb3/mb1433ui.htm slightly different chipset and lack of KB0, but the jumpers are all in the same places.
Following this guide i set everything for Intel DX4/100 - https://www.elhvb.com/webhq/models/486vlb3/mb1433ui.txt.html (of course assuming that "1" on all of them is on the right hand side, as only few jumpers on the board show where the first pin is. Unfortunately nothing has changed, the post card shows ----.
It does have some extra functionality available after pressing the button on the card, but nothing regarding the button was mentioned in the card's manual (which was super brief and only listed post codes)

Regarding the LEDs on the card, im getting +5V, +12V and -12V. +3.5V is not lit. CLK is lit. Reset lits up shortly when powering up and also when connecting reset jumper.

her
https://theretroweb.com/motherboards/?showIma … 67&cpuSocket1=5 - Try to find MBO here. There are 10 motherboard that looks like Your, but i bet they all have same jumper configuration. I didn't check....

Reply 11 of 18, by mkarcher

User metadata
Rank Oldbie
Rank
Oldbie
dinth wrote on 2022-11-18, 19:22:

Unfortunately nothing has changed, the post card shows ----.

This means that the processor doesn't successfully execute the first instructions of the BIOS.

dinth wrote on 2022-11-18, 19:22:

It does have some extra functionality available after pressing the button on the card, but nothing regarding the button was mentioned in the card's manual (which was super brief and only listed post codes)

I know that card. Nothing of the extra functionality is useful before you start getting POST codes at all.

dinth wrote on 2022-11-18, 19:22:

Regarding the LEDs on the card, im getting +5V, +12V and -12V. +3.5V is not lit. CLK is lit. Reset lits up shortly when powering up and also when connecting reset jumper.

This means reset (at least on the ISA bus) is working fine. This is a good thing. The missing "+3.3V" LED indicates -5V is missing (It's for +3.3 on PCI and -5 on ISA). This is likely due to your ATX power supply not supporting -5V. This might cause some boards to hold the processor in reset until all voltages are present, but as you see proper reset operation, the missing -5V supply is extremely unlikely to cause the issue you are observing. Please also check the IRDY LED. It should light up whenever the processor tries to access the BIOS. If IRDY doesn't light up, most likely the chipset doesn't receive any attempts from the processor to read the BIOS.

Do you use the same cache chips on all your boards? In that case, it is possible that one of the cache chips is damaged and takes down the 486 local bus. The cache is directly connected to the data pins
of the 486 processor on most boards, and might also be directly connected to some address pins (although I consider connecting 9 cache chips directly to the local bus without a buffer chip as bad design). Even without cache installed, you should get some POST codes if the board, the BIOS and the POST card are working.

Reply 12 of 18, by dinth

User metadata
Rank Newbie
Rank
Newbie

I cannot go back to testing two previous mobos because i have already disposed of them (sent them to a collector abroad who was willing to adopt them and even paid the postage).
-5V is definitely missing, but i always thought this may only cause some of the ISA soundcards not to work, so it was not my priority.

IRDY LED is interesting. Before i thought it doesnt light up at all, but now i have put in an AMD cpu (without changing any jumpers)
1) Most of the times it doesnt lit up at all.
2) Sometimes it lits and stays lit
3) Rarely it blinks for a tenth of a second.

Actually, during about 50+ tests this morning, on four different occasions i got POST codes on the display: C2, C3, C7 and 03

Reply 13 of 18, by mkarcher

User metadata
Rank Oldbie
Rank
Oldbie
dinth wrote on 2022-11-19, 09:10:

-5V is definitely missing, but i always thought this may only cause some of the ISA soundcards not to work, so it was not my priority.

Correct. That shouldn't be your priority right now.

dinth wrote on 2022-11-19, 09:10:

Actually, during about 50+ tests this morning, on four different occasions i got POST codes on the display: C2, C3, C7 and 03

OK, this looks like sometimes you get proper BIOS code execution. Boot-block based AWARD BIOSes start with C0-C5 or like that in the boot block and continue with 03 as soon as the decompressed actual BIOS is entered. This means that operation is unstable for some reason, but memory, processor, and ISA bus kind of work. Issues like this might be due to bad contacts or broken traces. Let's hope it's not a broken trace: Try contact cleaner on the BIOS chip and the CPU ("deoxit that socket" as some YouTubers like to say) and check whether things get better.

If bending the board slightly makes it work better/worse, it's likely a broken trace, though.

Reply 15 of 18, by dinth

User metadata
Rank Newbie
Rank
Newbie
dinth wrote on 2022-11-19, 09:10:

Actually, during about 50+ tests this morning, on four different occasions i got POST codes on the display: C2, C3, C7 and 03

OK, this looks like sometimes you get proper BIOS code execution. Boot-block based AWARD BIOSes start with C0-C5 or like that in the boot block and continue with 03 as soon as the decompressed actual BIOS is entered. This means that operation is unstable for some reason, but memory, processor, and ISA bus kind of work. Issues like this might be due to bad contacts or broken traces. Let's hope it's not a broken trace: Try contact cleaner on the BIOS chip and the CPU ("deoxit that socket" as some YouTubers like to say) and check whether things get better.

If bending the board slightly makes it work better/worse, it's likely a broken trace, though.
[/quote]

The thing is that it is AMI BIOS (chip dated 1993) and i cannot find those codes on AMI bios code lists

Reply 18 of 18, by CoffeeOne

User metadata
Rank Oldbie
Rank
Oldbie
dinth wrote on 2022-11-22, 18:58:
rasz_pl wrote on 2022-11-21, 23:37:

someone flashed wrong bios?

I dont think this is possible, all the mobos used to be in a working state lets say 20-25 years ago and nobody touched them since then.

Hello, please try to be less vague. ".... all the mobos used to be ..."
When you gave up on this mainboard now, then switch over to the next.
Make pictures, give info, whatever ....